As more AI-written texts flood the public space, people will feel an increasingly strong need to be certain that the texts they spend their time reading are written by actual humans. They will withhold trust and interest unless they are completely certain that they are reading human texts.
What people will look for is what we have here on the blog: the certainty that these are actual humans. And I don’t mean just me. It’s everybody who participates. Right now it seems like no big deal but soon enough it will be a mega big deal. Because all of a sudden we just won’t know. This will erode the social trust that is already at a low ebb.
Clarissa's Blog 
We have now seen examples of students using AI to write emails. I’m not entirely sure yet what I think about it. On one hand, student went through some effort to put together an email that didn’t look like a text message. On the other hand, it was very obvious to my colleague what happened since the AI simply made up stuff that doesn’t exist and the student confidently included it in the email (“AI hallucinations”). At the end, using AI didn’t help the student because he was not sophisticated enough to figure out what was real and what was not. Times are coming when a simple unsophisticated email with a few mistakes from a student will be preferable to a highly polished wall of text because at least we will know the person wrote it himself.
LikeLiked by 1 person
That’s true. Only people who have no other choice routinely eat McDonald’s. Nobody considers it a desirable dining option. The chic people eat fresh, seasonal, and prepared on the spot. In-person communication will be the “fresh, seasonal” equivalent of food. It will be a luxury good.
LikeLike
We may end up with the AI reading our emails to give us bullet point summaries and then having it reply the emails back with a wall of text. It will end up being AI-to-AI communication.
The science book publishers are now running book proposals and any submitted book chapters through a generative-AI text recognition checker as they are trying to avoid having books written by the AI. People definitely use AI in scientific writing and it took me by surprise (it really should not have). I did not figure out for myself yet what to do with the AI tools, so for the time being I am simply avoiding their use (for better or worse).
LikeLike
Sigh.
Every little outpost I’ve ever had on the internet, I’ve conscientiously replied to most comments, even people who seemed a little clueless, didn’t get it, whatever. I can see a time when we all just assume those are not real.
LikeLiked by 1 person
This is horrifying as someone who actually hand writes their stuff with a pen and notebook, I only use AI to write a description for stuff I sell on eBay. The idea of AI writing fiction is evil, fiction and academic papers should be written by humans. I’ve seen a lot of generic fiction on sites like Wattpad that might be written by AI, stuff about the nerdy girl who hooks up with the football player and they got trendy names. It’s too sophisticated and the grammar is too good for a human writer, I avoid these.
Conversely, I’ve read fan fictions on Wattpad that have a few misspellings and grammar errors that as fun and enjoyable to read as any professional published book. The authors are actual people who made a few mistakes but clearly are a big fan of their fandom and enjoyed writing their fiction, I’ve binge-read some of these in one afternoon and enjoyed them. I’d rather read a fan fiction by a talented, enthusiastic writer who makes a few spelling and grammar errors than a perfect AI book.
LikeLike
The future of the internet is tiny, gated or semi-gated communities (like this) where everyone knows each other. Because the vast proportion of the internet is going to be AI slop. It’s happening already.
LikeLiked by 1 person
This is a really interesting perspective. Thank you. Do you think this would largely apply to AI- and machine-translated texts as well? I don’t see why not, save for the tricky fact that those who obtain/purchase translations usually aren’t in the position to assess the quality of the end product (can’t speak/read the target language, or only do so at a surface level and have no idea how it comes off to a native speaker, what kind of feel and tone it generates). They just get dazzled by all the flashy vocabulary and the low/free price. So their spiffy, hyper-“accurate” translations might totally put off readers by sounding scammy, nauseatingly inauthentic, and interchangeable with all the competition spitting out similarly soulless AI texts, but they have no idea and won’t care until it affects their bottom line and someone makes them realize how much their texts suck.
LikeLike
I do almost exclusively medical, legal, or literary translations, and none of those can ever be AI-ed. But regular day-to-day translations with low importance and responsibility will definitely be machine translated. They are pretty much already.
LikeLike